Detecting hallucinations in large language models using semantic entropy(nature.com)

220 points by Tomte 2 years ago | 153 comments

Retrofitting language models to operate over bytes(nature.com)

99 points by theanonymousone 18 days ago | 20 comments

Open-access Nature research article describing a method to retrofit existing language models to operate directly over bytes instead of tokens.

Role play with large language models(nature.com)

83 points by m_kos 3 years ago | 64 comments

Large Language Models as Markov Chains(arxiv.org)

63 points by Anon84 1 year ago | 33 comments

Invalid SMILES are beneficial rather than detrimental to chemical lang models(nature.com)

3 points by kjhughes 2 years ago | 1 comments

Invalid SMILES beneficial rather than detrimental to chemical language models(nature.com)

20 points by Bluestein 2 years ago | 10 comments